Research map / Probabilistic and scientific modeling / Molecules and proteins
Cold-start recommendation models
Handles new users and items using dropout and Bayesian factorization models.
80 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Molecules and proteins, keyword cluster “decision · machine · statistical”
State History Tried Baselines Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
17 papers in 2023–24 and 21 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.5% to 2.3%.
Open questions the data points to
- Papers report FPMC-LR on Gowalla (P@5) anywhere from 0.048 to 0.116: comparisons against it depend on whose number you take.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · STELLAR: Spatial-Temporal Latent Ranking for Successive Point-of-Interest Recommendation (AAAI 2016 · 242 citations)
- 2017 · A Hybrid Collaborative Filtering Model with Deep Structure for Recommender Systems (AAAI 2017 · 285 citations)
- 2018 · Modeling Dynamic Missingness of Implicit Feedback for Recommendation (NeurIPS 2018 · 39 citations)
- 2019 · Doubly Robust Joint Learning for Recommendation on Data Missing Not at Random (ICML 2019 · 96 citations)
- 2020 · Sequential Latent Knowledge Selection for Knowledge-Grounded Dialogue (ICLR 2020 · 121 citations)
- 2021 · Uncovering Latent Biases in Text: Method and Application to Peer Review (AAAI 2021 · 27 citations)
- 2022 · A Dynamic Meta-Learning Model for Time-Sensitive Cold-Start Recommendations (AAAI 2022 · 15 citations)
- 2023 · Anytime User Engagement Prediction in Information Cascades for Arbitrary Observation Periods (AAAI 2023 · 4 citations)
- 2024 · Adversarial Socialbots Modeling Based on Structural Information Principles (AAAI 2024 · 15 citations)
- 2025 · Addressing Cold-Start Problem in Click-Through Rate Prediction via Supervised Diffusion Modeling (AAAI 2025 · 4 citations)
- 2026 · What Does Preference Learning Recover from Pairwise Comparison Data? (ICML 2026 · 0 citations)
Earliest papers here
- STELLAR: Spatial-Temporal Latent Ranking for Successive Point-of-Interest Recommendation (AAAI 2016 · 242 citations)
- Inferring a Personalized Next Point-of-Interest Recommendation Model with Latent Behavior Patterns (AAAI 2016 · 168 citations)
- Learning the Preferences of Ignorant, Inconsistent Agents (AAAI 2016 · 88 citations)
What has been tried
70 papers here report 621 measured changes (a component added, removed or swapped, compared with a baseline): helped 74%, no clear effect 9%, hurt 13%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Remove user filtration | Homogeneous social networks · policy-learning quality | Hurt | Adversarial Socialbots Modeling Based on Structural Information Principles (AAAI 2024 · 15 citations) | |
| use the CGC architecture instead of multi-task MLP layers | Ds A · AUC | 0.697±0.001 -> 0.710±0.002 | Helped | Backdoor Adjustment via Group Adaptation for Debiased Coupon Recommendations (AAAI 2024 · 9 citations) |
| Add an MDN to refine GT1 predictions | Real testing data · NLL | 6.795 -> 6.626 | Helped | Predicting Real-World Penny Auction Durations by Integrating Game Theory and Machine Learning (AAAI 2024 · 6 citations) |
| Add capability embedding to helping policy training | Smart Help evaluation · SR | 0.374 -> 0.419 | Helped | Smart Help: Strategic Opponent Modeling for Proactive and Adaptive Robot Assistance in Households (CVPR 2024 · 5 citations) |
| Add dropout to the diffusion model | MovieLens-1M Cold, Warm-a, Warm-b, Warm-c · AUC | 0.7456 vs 0.7456; 0.7980 vs 0.7980; 0.8059 vs 0.8059; 0.8091 vs 0.8090 | No clear effect | Addressing Cold-Start Problem in Click-Through Rate Prediction via Supervised Diffusion Modeling (AAAI 2025 · 4 citations) |
| Add zero-inflation to the hold checkout model | NYPL checkout prediction · balance of hold and browser checkout prediction performance | Helped | Identifying and Addressing Disparities in Public Libraries with Bayesian Latent Variable Modeling (AAAI 2024 · 4 citations) | |
| Use multiple checkpoints instead of a single checkpoint | FTRACE-TREx, candidate set 15K · MRR | 0.1658 -> 0.1596 | Hurt | The Mirrored Influence Hypothesis: Efficient Data Influence Estimation by Harnessing Forward Passes (CVPR 2024 · 3 citations) |
| add latent-space RL exploration | ParisAttack2015, IC, 20% budget · influence spread | 515237.59 -> 637621.59 | Helped | REM: A Scalable Reinforced Multi-Expert Framework for Multiplex Influence Maximization (AAAI 2025 · 2 citations) |
| Use the rescaled rather than unscaled confidence-bound procedure | Synthetic sub-interval certification · power | Helped | Statistical Inference for Fairness Auditing (NeurIPS 2024 · 2 citations) | |
| Use BayesCNS rather than the non-behavioral approach | Stationary simulated environment, across predictive power settings · CTR | Helped | BayesCNS: A Unified Bayesian Approach to Address Cold Start and Non-Stationarity in Search Systems at Scale (AAAI 2025 · 1 citation) | |
| add Counterfactual Sequence Augmentation | KuaiRec · MRR | 0.476 -> 0.512 | Helped | Counterfactual Task-augmented Meta-learning for Cold-start Sequential Recommendation (AAAI 2025 · 1 citation) |
| Add LCL-style cross-feature context effects to CRCS | Hotels · NLL | 536 -> 536 | No clear effect | Preference Learning of Latent Decision Utilities with a Human-like Model of Preferential Choice (NeurIPS 2024 · 1 citation) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| FPMC-LR | Gowalla | P@5 | 0.048 – 0.116 | 2 papers, e.g. AAAI 2016 and AAAI 2016 |
What reviewers will expect
Datasets and baselines named in the 16 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- LPMC · 2 of 16 papers
Baselines
- Wide & Deep · 2 of 16 papers
- DropoutNet · 2 of 16 papers
- SASRec · 2 of 16 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 38 papers; from OpenAlex and the papers' affiliations).
- Carnegie Mellon University · 2 papers
- Beijing Institute of Technology · 2 papers
- Google · 2 papers
- Stanford University · 2 papers
- University of California, Berkeley · 2 papers
Use it
Download BibTeX of 40 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Prediction-powered and conformal inference (129 papers)
- Statistical analysis of ML practice (122 papers)
- Machine learning for physical simulation (96 papers)
- Counterfactual and statistical explanations (81 papers)
- Sparse concepts and hypothesis testing (78 papers)
Elsewhere in the topic
- Generative protein structure design (142 papers)
- Neuro-symbolic probabilistic layers (120 papers)
- Structure-based 3D molecule generation (107 papers)
- Sparse graphical and tensor models (97 papers)
- Flow matching generative modeling (97 papers)
- Multimodal trajectory prediction (95 papers)
- Probabilistic embeddings and relational inference (93 papers)
- Equivariant and disentangled latent representations (90 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).